Industrialised & Responsible AI

Driving future business performance 

In today's rapidly evolving business landscape, Sopra Steria is convinced that the widespread adoption of Artificial Intelligence is no longer just a strategic advantage — it's a necessity. It is the lynchpin of future business performance. 

Managing the effective use of AI throughout a large corporation poses significant challenges, which is why we seek to support companies throughout every step of their AI deployment, from initial idea, through proof of concept up to fully-fledged deployment at scale. 

Industrialised rollout of AI across the enterprise business lines necessitates extensive technological and regulatory compliance and therefore companies must commit to responsible and ethical AI principles. 

That is why Sopra Steria is a key investor in Confiance.ai, a unique French community dedicated to the design and industrialisation of trustworthy critical systems based on artificial intelligence. 

Alongside Confiance.ai, we aim to establish a clear and transparent methodology allowing business to qualify the trustworthiness of data-based intelligent systems that companies can integrate into industrial products and services, safe in the knowledge of their reliability. 

This covers aspects such as explainability, robustness, monitoring, uncertainty quantification, synthetic data generation, and trustworthy learning. 

Combining the technological expertise derived from Confiance.ai with our expertise in AI governance, we help our customers assess, validate and deploy responsible AI at scale while also ensuring they comply with all necessary regulatory requirements.

What is Responsible AI? 

As AI systems become increasingly autonomous and capable of making decisions that impact individuals and societies, it becomes imperative to ensure that their deployment is guided by principles that prioritise fairness, transparency, accountability, and human well-being. 
 
Responsible AI promotes equitable treatment, explains algorithmic decisions, and establishes mechanisms for oversight and accountability. By prioritising ethical considerations, it aims to mitigate biases and ensure trustworthiness in AI systems. 

Whitepaper

Responsible artificial intelligence

As organisations race to seize AI’s benefits, prioritising responsibility is key. Embracing responsible AI practices is not just about staying ahead but building a sustainable competitive advantage. 

Read more

Why industrialise responsible AI deployment? 

  • Scalability: Industrialising AI rollout allows companies to scale up implementations across various business units and processes, ensuring widespread adoption and maximising the benefits of AI technologies.
  • Efficiency: Standardising AI implementations streamlines processes and reduces duplication of efforts, leading to increased operational efficiency and cost savings.
  • Competitive Advantage: Rapid industrialisation of AI enables companies to stay ahead of competitors by leveraging advanced technologies to improve decision-making, optimise workflows, and deliver superior products and services.
  • Innovation acceleration: Industrialised AI facilitates rapid experimentation and iteration, fostering a culture of innovation within organisations and enabling the development of cutting-edge solutions to complex business challenges.
  • Regulatory compliance: New legislation, such as the European Union's AI Act, oblige companies to adapt their AI governance to mitigate risks associated with bias, privacy, and security. Companies must be fully compliant within three years.

AI industrialisation challenges

Data Quality & Accessibility

Challenge: Inconsistent or poor-quality data across different departments or sources can hinder AI deployment.

Solution: Establish data governance to ensure data quality, accessibility, and consistency. Invest in data cleaning and preprocessing to enhance dataset reliability.

Talent & Skills Gap

Challenge: Shortage of skilled professionals with expertise in AI development, deployment, and management.

Solution: Invest in training programs to upskill existing employees and attract top AI talent. Foster collaboration between data scientists, engineers, and domain experts.

Infrastructure & Integration Complexity

Challenge: Integrating AI systems into existing infrastructure can be complex and time-consuming.

Solution: Invest in flexible and scalable infrastructure solutions such as cloud computing. Prioritise interoperability when selecting AI tools and platforms.

What is the AI Act? 

The European Union’s AI Act is the world’s first comprehensive global regulatory framework governing AI development, deployment, and use within the bloc. It categorises AI systems by risk level and imposes corresponding obligations. High-risk systems face stringent regulations, while transparency requirements apply to lower-risk systems. 

Compliance is vital, enabling companies to demonstrate their ethical added value and to reassure end users, not to mention that failing to comply can lead to substantial penalties. These can range from €10 million to €40 million or 2% to 7% of global annual turnover. Proactive assessment and measures enable companies to ensure compliance while upholding ethical standards and user safety. 

The EU’s AI Act: How companies turn compliance into competitive edge
Read more

How Sopra Steria helps 

With profound knowledge and experience in industrialisation of responsible AI, Sopra Steria will help you achieve your digital transformation to enable you to reach your maximum potential. 

Expertise 

Guidance implementing responsible AI practices, drawing on our vast experience across all key verticals. 

Framework development 

Construct frameworks to integrate Responsible AI principles into a company’s AI strategy. 

Ethical AI design 

Conceptualise and realise AI systems with built-in ethical considerations to mitigate biases and ensure responsible decision-making. 

Compliance 

Audit and assessment of regulatory compliance and implement measures to address any compliance gaps.

Training and education 

Development of training programs to empower employees with the skills to understand and implement responsible AI practices. 

Test and learn 

Ongoing monitoring and evaluation of AI systems to identify and address ethical issues or risks, ensuring continuous improvement. 

Our offer

Discover how AI is helping our clients

How Norad and Sopra Steria leverage AI and cloud tech to fight child illiteracy

Jul 11, 2024, 17:21 PM
Title : How Norad and Sopra Steria leverage AI and cloud tech to fight child illiteracy
Button label :
A joint project by the Norwegian Agency for Development Cooperation and Sopra Steria leverages AI and cloud tech to boost literacy by creating open, accessible education resources for some of the world’s poorest children. 

 

Only one in ten 10-year-olds in sub-Saharan Africa can read a simple sentence, despite many having completed several years’ schooling. 

That translates to some 617 million children and adolescents not achieving minimum proficiency levels in reading and mathematics, according to UNESCO estimates. 

However, an ambitious project spearheaded jointly by the Norwegian Agency for Development Cooperation (Norad) and Sopra Steria seeks to combat this child illiteracy by leveraging the power of AI and cloud computing to create a huge range of open and accessible educational resources. 

The Global Digital Library (GDL) uses tech to solve real-world challenges by translating and creating new educational assets in children’s mother tongue.

The project seeks to align with UN Sustainability Goal number 4, ensuring inclusive, equitable and quality education for all and improving basic reading, writing and arithmetic. 

The goal of the Global Digital Library is to provide children with learning resources in a safe, open, and inclusive digital platform, translated and available in both majority and minority languages. 

AI creating new educational resources 

The Global Digital Library is a collaborative project with global actors including Unesco, Unicef, The Global Book Alliance and Creative Commons. 

The Library contains more than 11,000 reading and math resources, translated into more than 140 minority and majority languages because we know children learn best in their mother tongue. 

The project provides more children and young people with access to learning resources in a language they use and understand. 

The content is open source, interoperable, and available to everyone, and is designed for teachers, parents, and publishers, and most importantly, children. It is accessible on all mobile devices and can be used in offline mode. 

In addition to children's books, GDL now also has mathematics games and videos for children aged 6 to 12 based on content from Khan Academy. 

The content is created by Sopra Steria employees with educational and mathematical backgrounds. Many of the games have great illustrations that allow children to absorb the content through text and visuals. 

600 million children and young people worldwide lack basic skills in reading and maths. The GDL uses AI and cloud tech to create open, accessible education resources in children’s mother tongue. 

Huyen Tran Titti Ho Brekken

Senior Adviser, Department for Knowledge and Innovation, Norad

Strategic marketing to raise GDL profile 

Sopra Steria also developed a web platform to host all learning resources and uses artificial intelligence to translate and produce new assets. 

Sopra Steria is also boosting the Library’s accessibility, and visibility through strategic marketing, direct contact with key partners and workshops, and their software engineers have created interactive games to make learning more fun. 

The project has also worked on reading training resources for children who cannot read, initially in English. These will be published in 2024. 

The project collaborates with partners to translate books and math games into different languages but also to increase the use of resources. 

Today, the Library has more than 670,000 users, mainly across Africa and Asia, and this is just the beginning. 
Cloud tech to create a sustainable digital library 

Although the Library has content in some 140 languages, it has the capacity to accommodate content in some 300 languages. 

Last year, the "Translate a Story" translation campaign led to the production of more than 700 children's books translated into 12 Ghanaian and 29 Indonesian languages. 

This was done in close collaboration with Unesco teams in France, Ghana and Indonesia, and their respective governments. 

Next, the focus is on the use of resources. Our pilot school programme, where resources will be tested over a minimum 6-week period in close collaboration with teachers, launched in South Africa, Ethiopia, and Vietnam. 

The goal is a sustainable library that can be maintained and operated by partners, schools, and teachers. Assets are designed to be used in schools and in children’s free time, in libraries, and at home. 

Test-and-learn platform methodology 

Regular feedback from teachers and students using the resources will give us opportunities to continuously improve and develop the resources. Teachers will also be able to create their own resources for the library. 

Sopra Steria developers improved the GDL platform, adapting H5P interactivity and facilitating download and sharing solutions. We use a test-and-learn approach to implement feedback from users, and in collaboration with designers and content teams, we work to make the platform as user-friendly as possible and tailored to the target audience. 

Insight, illustrations, and landing pages UX designers create flow and display both on the platform and in resources based on regular user tests conducted by children and youth in Africa and Asia. We have contributed thousands of illustrations used on the website, landing pages, and social media campaigns. 

The target audience is far from Norway geographically and, in many cases, culturally, and being close through conversations, interviews, observation, and user testing has provided valuable insight into the development process. 

Data-informed marketing strategy 

Our marketing specialists have been running data-driven marketing campaigns targeting the markets of Rwanda, Ghana, and Kenya for several years. The goal has been to reach parents and teachers to create awareness, traffic, and boost platform use. With a data-driven approach, we build strategies based on data and insight. 

Sopra Steria has produced videos and dynamic ad content with visual material for campaigns such as Hey teacher! and Hey parent!, using different formats and ad placements that play on the strengths and characteristics of each location in channels like Google Ads, Facebook, Instagram, Display, and YouTube. 

Numbers and statistics are collected and analysed in Google Looker Studio so we can adjust Key Performance Indicators (KPIs) and assess success based on the goals set for each campaign. 

In addition, campaigns are synchronised to run alongside relevant international educational conferences with geographical targeting. For example, we set dedicated social campaigns alongside Digital Learning Week in Paris and the MEducation Alliance in Washington, to build brand awareness for the platform for possible partners. 

Sopra Steria also produces content for weekly posting on social media, X (Twitter), and LinkedIn, to highlight content in the platform. 

We want more users, more reach, and more content, especially games and interactive learning resources. The potential for tech to contribute is huge. 

Tech offers an unprecedented chance to democratise education. Because in the end, every child deserves the right to learn, regardless of where they live. 

  • 11,000 reading and maths resources on the GDL
  • 140 different languages currently available on the platform
  • 670,000 regular users of the platform, mainly spread across Asia and Africa

We’re delighted that our AI and cloud insight has had such a profound, transformative and positive impact on this critical education project. 

Cecilie Eftedal

Content Lead, Sopra Steria Norway

Tags :